Markov Chain Models for Predicting Student Progression and Dropout in Nigerian Universities

This study employs Markov chain models to analyse and predict student academic progression and dropout rates in Nigerian universities, using longitudinal enrolment and progression data from a federal university in the South South geopolitical zone as a case study. Student attrition is a significant challenge facing Nigerian higher education institutions, with dropout rates affecting both institutional accreditation standings and national human capital development goals. The study conceptualises each year of academic study as a state in a discrete-time Markov chain, with additional absorbing states representing graduation and dropout. Transition probability matrices are estimated from five academic cohorts enrolled between 2016 and 2020, covering 2,400 students across the faculties of sciences, social sciences, and humanities. Stationary distribution analysis is conducted to derive long-run estimates of graduation and dropout probabilities, and sensitivity analyses are performed to assess how changes in year-to-year transition probabilities affect steady-state outcomes. Results reveal that the highest dropout hazard is concentrated in the second and third years of study, a pattern consistent with the academic and financial pressures documented in recent Nigerian higher education research. The study further identifies faculty of enrolment and gender as moderating variables, with female students in science programmes exhibiting lower completion probabilities than their male counterparts. Policy implications for early academic intervention and scholarship targeting are discussed. Keywords: Markov chains, student progression, dropout, transition probabilities, Nigerian universities